Electrochemical Migration Testing

Electrochemical Migration Testing

Quality & Inspection Standards

PRODUCTS CENTER

Get Custom Quote

PRODUCTS CENTER

Supported formats: PDF, DWG, Gerber, Excel (Max 50MB)
Submit
ECM Testing IPC‑TM‑650 Methods 2.6.14 / 2.6.14.1

Electrochemical Migration Testing – ECM vs CAF, IPC Methods & Design Prevention

Electrochemical migration (ECM) is the surface‑based cousin of CAF – and it is equally destructive, but much faster to appear. While CAF grows inside the laminate over months, ECM forms dendritic metal crystals on the PCB surface under high humidity and bias. We have seen silver dendrites bridge 0.3mm gaps in less than 48 hours at 85°C/85%RH with 50V bias. The result: intermittent shorts that disappear when the board dries, making field diagnosis nearly impossible. ECM is driven by surface contamination, exposed copper or silver, and inadequate solder mask coverage. Per IPC‑6012E §3.9.2, resistance to electrochemical migration is a requirement for solder mask and surface finishes. This guide covers ECM test methods (IPC‑TM‑650 2.6.14, 2.6.14.1), the difference between ECM and CAF, critical design parameters (spacing, voltage, surface finish), and how to qualify materials and processes for high‑humidity environments.

Published: May 15, 2026  |  10 min read  |  #ECM #ElectrochemicalMigration #DendriticGrowth #HumidityReliability

1. ECM vs CAF – Surface Migration vs Internal Filament

Engineers often confuse ECM and CAF because both involve electrochemical transport of metal ions. The differences are critical for root cause analysis.

Characteristic ECM (Surface Migration) CAF (Internal Filament)
Location On the PCB surface (solder mask, exposed pads) Inside the laminate along glass fibers
Metal source Surface finish (Ag, Sn, Cu, ENIG Ni) Copper from PTH barrels
Appearance Dendritic, branching, metallic crystals Dark, thread‑like along glass bundle
Detection Visible under 50‑200x stereo microscope Requires microsection (200‑500x)
Time to failure Hours to days (high humidity) Months to years
Primary driver Ionic contamination + exposed metal + humidity Halide ions + glass/resin interface + humidity
IPC test method 2.6.14, 2.6.14.1 (solder mask ECM) 2.6.25

Table 1: Key differences between ECM and CAF. ECM is surface‑visible, CAF requires sectioning.

In a satellite communication product, silver‑finished pads spaced 0.4mm apart grew dendrites after 500 hours of 85°C/85%RH testing. The dendrites were visible at 50x. CAF was not present. The root cause was a combination of exposed silver (solder mask misregistration) and ionic residues from a no‑clean flux. Changing to ENIG and improving solder mask registration eliminated ECM.

2. ECM Test Methods – IPC‑TM‑650 2.6.14 & 2.6.14.1

IPC‑TM‑650 defines two ECM test methods. Both use comb patterns and measure insulation resistance over time under humidity/bias.

  • 2.6.14 – Electrochemical Migration Resistance of Solder Mask: Tests the solder mask's ability to prevent ECM between adjacent conductors. The comb pattern is coated with solder mask, then subjected to 85°C/85%RH with 50V bias. Duration: 168‑500 hours. SIR must remain ≥1×10⁸ Ω, and no dendritic growth visible at 50x.
  • 2.6.14.1 – Electrochemical Migration Test (General): Similar to 2.6.14 but can be applied to any surface finish or laminate. Often used to compare different surface finishes (ENIG vs immersion silver vs OSP).

Typical test setup:

  • Comb pattern with defined line/space (e.g., 0.3mm, 0.4mm, 0.5mm).
  • Temperature: 85°C ± 2°C.
  • Relative humidity: 85% ± 3% RH.
  • Bias voltage: 50V DC (or lower for low‑voltage products).
  • Measurement intervals: 24, 48, 96, 168, 336, 500 hours.
  • Pass/fail: SIR ≥1×10⁸ Ω and no visible dendrites at 50x.

In one qualification, an immersion silver finish failed ECM after 96 hours – silver dendrites grew across 0.3mm gaps. The same board with ENIG passed 500 hours. The customer switched to ENIG for all high‑humidity applications despite a 20% cost increase.

3. Critical Factors – Metal Type, Spacing, Humidity, Bias, Contamination

The ECM time‑to‑failure (TTF) is governed by multiple variables. The table below shows relative susceptibility (industry data).

Factor Low Risk High Risk Typical TTF at 85/85
Surface finish ENIG (Ni/Au), OSP Immersion silver (Ag), HASL (Sn) Ag: 50‑200 hours; ENIG: >500 hours
Conductor spacing ≥0.5mm ≤0.3mm TTF ∝ spacing²
Bias voltage ≤12V ≥50V TTF ∝ V⁻²
Relative humidity ≤50% RH ≥85% RH TTF ∝ RH⁻³ (very strong)
Ionic contamination ROSE <0.5 µg/cm² ROSE >1.5 µg/cm² Halides accelerate migration
Solder mask coverage Full coverage, no exposed traces Exposed copper or pads Exposed metal dramatically increases risk

Table 2: ECM risk factors. Silver and tin are the most problematic surface finishes.

In a power supply design with immersion silver finish and 0.35mm spacing between high‑voltage traces (48V), ECM failures occurred after 2 weeks of field operation in a humid environment (Singapore). The fix: change to ENIG and increase spacing to 0.5mm. The board passed 1000 hours of ECM testing.

4. Failure Modes – Silver Migration, Tin Whiskers, Copper Dendrites

Different metals produce different ECM morphologies and risks.

  • Silver migration (most common): Silver ions dissolve in moisture, then plate out as dendritic crystals on the cathode. The dendrites are thin, branching, and can grow across 0.5mm gaps in hours. Immersion silver finishes are most susceptible. Prevention: use ENIG or OSP instead, or encapsulate silver with solder mask.
  • Tin migration (HASL, tin‑lead): Tin whiskers are a form of ECM but can also grow without humidity (stress‑induced). Tin migration is slower than silver but still dangerous. Alloying with lead reduces risk (but lead is restricted). Pure tin finishes require conformal coating.
  • Copper dendrites: Grow when copper is exposed (scratched solder mask, unplated vias). Copper dendrites are thick and can carry significant current, causing hard shorts.
  • Nickel migration (ENIG): Rare – nickel is less electrochemically active. This is why ENIG is preferred for high‑reliability.

In a medical device, a 0.25mm gap between an exposed copper pad and a nearby via grew copper dendrites after 72 hours of 85/85 testing. The copper pad had a solder mask misregistration of 0.1mm. Fixing the mask registration eliminated the problem. The lesson: do not leave any copper exposed on biased nets.

5. Acceptance Criteria – SIR Drop, No Visible Dendrites, IPC Limits

IPC‑6012E §3.9.2 states: "Solder mask shall be resistant to electrochemical migration when tested per IPC‑TM‑650 2.6.14." However, specific numeric limits are not given. Industry practice defines:

  • SIR after 168 hours: ≥1×10⁸ Ω (100 MΩ) for most applications. Aerospace often requires ≥1×10⁹ Ω.
  • SIR after 500 hours: ≥1×10⁷ Ω (10 MΩ) minimum, but preferably >1×10⁸ Ω.
  • Visual inspection at 50‑100x: No dendritic growth, no discoloration, no corrosion products.
  • Change in resistance: A drop of more than one order of magnitude (e.g., from 1×10¹⁰ to 1×10⁸) is a process warning, even if still above the pass limit.

In one automotive program, the customer required no visible dendrites at 50x after 500 hours of 85/85 with 50V bias. A batch of boards with immersion silver finish failed – dendrites were visible at 80 hours. The supplier switched to ENIG, and the next batch passed 1000 hours. The lesson: visual inspection is as important as electrical measurement.

6. Design Prevention – Minimum Spacing, Solder Mask Overlap, Conformal Coating

ECM can be prevented by design rules, not just material selection.

  • Minimum conductor spacing: For biased nets exposed to humidity, use ≥0.5mm spacing for <50V. For 50‑100V, use ≥0.8mm. This is the single most effective prevention.
  • Solder mask overlap: Solder mask should cover all traces and pads except where components are soldered. Mask registration tolerance must be such that no copper is exposed on biased nets. A typical rule: mask clearance ≤0.05mm from pad edge.
  • Avoid immersion silver for high‑humidity environments: Silver is highly susceptible to migration. If silver is required (e.g., for press‑fit connectors), ensure spacing ≥0.8mm or apply conformal coating.
  • Conformal coating: For products operating in >70% RH environments, conformal coating (acrylic, silicone, urethane) is highly effective at preventing ECM. Coating thickness ≥25µm (1mil) is typical.
  • Cleanliness: Ionic contamination <1.0 µg/cm² (ROSE) and chloride/bromide <0.3 µg/cm² (IC). Clean boards are much less likely to migrate.

In an outdoor telecom product, the original design had 0.4mm spacing and no conformal coating. ECM failures occurred after 6 months. The redesigned board increased spacing to 0.6mm and added silicone conformal coating. No failures in 5 years. The cost increase was 5% – far cheaper than field service calls.

ECM Prevention Checklist

  1. Identify all biased nets operating in humid environments. Any net with voltage >5V in >50% RH is a candidate for ECM.
  2. Increase spacing to ≥0.5mm for <50V. For higher voltage, use IPC‑2221 creepage tables.
  3. Avoid immersion silver for high‑reliability / high‑humidity. Prefer ENIG or OSP.
  4. Ensure solder mask covers all copper edges. Mask misregistration should not expose biased copper.
  5. Specify ECM testing per IPC‑TM‑650 2.6.14 or 2.6.14.1 on first article. Use comb pattern matching minimum spacing.
  6. For products with >70% RH exposure, apply conformal coating. Acrylic (AR) is common; silicone (SR) for high temperature.
  7. Maintain ionic cleanliness ROSE <1.0 µg/cm², chloride/bromide <0.3 µg/cm².

7. FAQ

Q: What is the difference between ECM and CAF?

A: ECM occurs on the PCB surface, forming dendritic crystals visible under a microscope. CAF occurs inside the laminate along glass fibers and requires microsection to detect. ECM is typically faster (hours to days); CAF takes months to years.

Q: Why is immersion silver so susceptible to ECM?

A: Silver ions have high mobility in water and low reduction potential, meaning they plate out easily. Even trace amounts of silver on the surface (from incomplete cleaning) can initiate dendrites. ENIG uses nickel (low migration) and gold (inert) – much more stable.

Q: Does IPC require ECM testing for all Class 3 boards?

A: No – IPC‑6012E §3.9.2 requires solder mask to be resistant to ECM when tested per 2.6.14, but it does not mandate the test for every board. It is typically performed as part of material qualification (solder mask, surface finish) and first article. Many high‑reliability customers require it for products used in humid environments.

Q: Can conformal coating stop ECM completely?

A: Yes – if applied properly (no voids, minimum thickness 25µm). Conformal coating physically isolates the conductors, preventing the moisture film needed for ion transport. However, if the coating has pinholes or is damaged, ECM can still occur locally.

Concerned about electrochemical migration in your high‑humidity application? Our engineers will review your design, recommend minimum spacing, surface finishes, and ECM test plans per IPC‑TM‑650 and IPC‑6012.

Request ECM Risk Assessment →

References: IPC‑TM‑650 2.6.14, 2.6.14.1; IPC‑6012E §3.9.2; IPC‑2221B (creepage tables); J‑STD‑004 (flux residues). IPC‑TM‑650 Official Method.

Get Custom Quote

PRODUCTS CENTER

Supported formats: PDF, DWG, Gerber, Excel (Max 50MB)
Submit